Measurable reconstruction from a dense trace #
A continuous real function is reconstructed Borel-measurably from its values on a countable dense subset, with the zero function on the branch where no continuous reconstruction exists.
theorem
BGW.Probability.measurable_continuous_reconstruction
{X : Type u_1}
{Ω : Type u_2}
[TopologicalSpace X]
[MeasurableSpace Ω]
[PolishSpace C(X, ℝ)]
(D : Set X)
[Countable ↑D]
(hD : Dense D)
(a : Ω → ↑D → ℝ)
(ha : Measurable a)
:
Measurable fun (ω : Ω) => if h : ∃ (g : C(X, ℝ)), ∀ (t : ↑D), g ↑t = a ω t then Classical.choose h else 0
Reconstruction of a continuous function from a countable dense trace is Borel.